i18n: add Simplified Chinese (zh-CN), all 1307 strings

[xwnxzz]

Adds locales/zh-CN.json with the full catalogue translated (1307 entries), registers it in
src/i18n/registry.ts, and adds a 简体中文 section to locales/CREDITS.md — following
locales/README.md exactly: keys untouched, {placeholders} untouched, plural categories kept,
one line in the registry.

Checks run

npm run i18n:check   → zh-CN: 1307 entries, 0 missing, 0 extra, 0 placeholder mismatches
npm run i18n:lint    → no hardcoded UI strings found
npm run build        → tsc --noEmit && vite build: OK
npx vitest run       → 2116 passed, 3 failed

The 3 failures (cjkTypography, sketch/textDrag, sketch/trimRawCursor) also fail on the
unmodified tree on this machine, so they are not caused by this change.

One thing worth your attention (pre-existing, not touched here)

Several tests assert English label text. On a machine whose Node/ICU default locale is zh-CN,
detectLocale() now resolves to zh-CN and 101 further tests fail purely because the strings
they assert are translated. On an English-locale runner (CI) they all pass. So the suite is
locale-sensitive today; pinning the locale in a test setup file would fix that, but it is your call
and this PR deliberately does not touch tests.

Notes

• "zh-CN" is registered so an exact navigator.language of zh-CN is auto-detected. A bare zh
  still falls back to English (the picker offers 简体中文).
• 26 strings warn as "identical to English": symbols and brands (D, R, B / I, TinkerAtlas,
  {a} ∩ {b}, …) where translating would be wrong.
• Nothing outside locales/ and the one registry line is touched.